The relationship of Minnesota Multiphasic Personality Inventory (MMPI) profile clusters to pain behaviors

Establishing Central Sensitization-Related Symptom Severity Subgroups: A Multicountry Study Using the Central Sensitization Inventory

OBJECTIVES

The goal of this study was to identify central sensitization-related symptom severity subgroups in a large multicountry sample composed of patients with chronic pain and pain-free individuals using the Central Sensitization Inventory (CSI).

METHODS

A large, pooled international (N = 8 countries) sample of chronic pain patients plus healthy subjects (total N = 2,620) was randomly divided into two subsamples for cross-validation purposes. First, a hierarchical cluster analysis (HCA) was performed using CSI item-level data as clustering variables (test sample; N = 1,312). Second, a latent profile analysis (LPA) was conducted to confirm the optimal number of CSI clusters (validation sample; N = 1,308). Finally, to promote implementation in real-world clinical practice, we built a free online Central Sensitization Inventory Symptom Severity Calculator.

RESULTS

In both HCA (N = 1,219 valid cases) and LPA (N = 1,245 valid cases) analyses, a three-cluster and three-profile solution, respectively, emerged as the most statistically optimal and clinically meaningful. Clusters were labeled as follows: (i) Low Level of CS-Related Symptom Severity, (ii) Medium Level of CS-Related Symptom Severity, and (iii) High Level of CS-Related Symptom Severity.

CONCLUSIONS

Our results indicated that a three-cluster solution clearly captured the heterogeneity of the CSI data. The calculator might provide an efficient way of classifying subjects into the cluster groups. Future studies should analyze the extent to which the CSI cluster classification correlates with other patient-reported and objective signs and symptoms of CS in patients with chronic pain, their associations with clinical outcomes, health-related costs, biomarkers, (etc.), and responsiveness to treatment.

Prospective Comparison of the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) and MMPI-2-Restructured Form (MMPI-2-RF) in Predicting Treatment Outcomes Among Patients with Chronic Low Back Pain

The purpose of the current study was to examine the relative utility of the most updated MMPI adult instrument, the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF), which was designed to address psychometric limitations of the MMPI-2. To this end, we compared mean scores and correlates of emotional distress treatment outcomes using the Depression Anxiety Stress Scales in a sample of 230 patients (73 males, 157 females) who had completed an interdisciplinary chronic pain rehabilitation program. Structural equation modeling analyses indicated that higher scale scores from all the MMPI-2-RF substantive domains were meaningfully associated with worse emotional distress outcomes, whereas the MMPI-2 Clinical Scales generally did not have any meaningful associations. Similar results were found in additional analyses using a clinically significant change framework with more direct clinical implications. The results of this study provide preliminary support for the use of the MMPI-2-RF among patients with chronic low back pain.

Pain threshold reflects psychological traits in patients with chronic pain: a cross-sectional study

Background

Chronic pain enhances sensory sensitivity and induces the biased development of psychological traits such as depression and pain catastrophizing, leading to the formation of heterogeneous conditions. Fluctuations in the sensory-related thresholds of non-injured sites (with normal peripheral tissue) in patients with chronic pain are thought to be related to central sensitization. The objectives of this study were to analyze the association between pain tolerance thresholds (PTTs) in non-injured sites and the psychological traits of patients with chronic pain and to evaluate the usefulness of PTT measures in assessments of pathological conditions related to chronic pain.

Methods

This study included 57 patients with chronic pain. The PTTs were measured in non-injured sites with quantitative sensory testing (QST) with electrical stimulation and then classified with cluster analysis. The Short-Form McGill Pain Questionnaire was used to subjectively assess pain in the injured sites. The Minnesota Multiphasic Personality Inventory (MMPI) was used to assess the patients’ psychological traits.

Results

Based on the cluster analysis of PTTs, the patients were classified into a High-Sensitivity group and an Others group consisting of the remaining patients. The results of the MMPI profiles showed that the High-Sensitivity group included significantly more patients with the Neurotic Triad pattern and no patients with the Conversion V pattern. The scores of the hypochondriasis and hysteria scales were significantly lower in the High-Sensitivity group than in the Others group.

Conclusions

This study indicated that patients with chronic pain can be classified according to PTTs in non-injured sites and suggests that patients with High-Sensitivity have characteristic psychological traits. Assessment of PTTs in non-injured sites would be useful for evaluating the psychological condition of patients with chronic pain.

Early psychologically informed interventions for workers at risk for pain-related disability: does matching treatment to profile improve outcome?

PURPOSE

This randomized controlled trial had two main aims. The first aim was to investigate the effect of early preventive, psychologically informed, interventions for pain-related disability. The second aim was explore whether people who are matched to an intervention specifically targeting their psychological risk profile had better outcomes than people who were not matched to interventions.

METHODS

A total of 105 participants were recruited from their workplace, screened for psychological risk factors and classified as being at risk for long-term pain-related disability. They were subgrouped into one of three groups based on their psychological profile. Three behaviorally oriented psychological interventions were developed to target each of the three risk profiles. Half of the participants were assigned a matched intervention developed to target their specific profile, and half were assigned an unmatched intervention. After treatment, repeated measure ANOVAs and v2 tests were used to determine if treatments had an effect on primary and secondary outcomes including perceived disability, sick leave, fear and avoidance, pain catastrophizing and distress, and if matched participants had better outcomes than did unmatched.

RESULTS

Treatments had effects on all outcome variables (effect sizes d ranging between 0.23 and 0.66), but matched participants did not have better outcomes than unmatched.

CONCLUSIONS

Early, preventive interventions have an impact on a number of outcome variables but it is difficult to realize a matching procedure. More in-depth research of the process of matching is needed.

Psychometric methods for measuring pain

Pain is a complex, multidimensional construct that includes a number of domains, such as pain severity, sensory qualities of pain, pain beliefs, coping strategies, and the functional impact of pain. In addition a variety of psychosocial factors, including affective state, the subjectivity of pain, and the degree to which one uses pain complaints to communicate about other psychosocial problems, strongly influence people's report of pain and disability. Thus it is important to recognize the various dimensions that contribute to the construct of pain and to consider evaluation methods that accurately provide for assessment of each of these dimensions. In some cases this would require administration of several domain-specific measures. Alternatively use of multidimensional measures is a viable option that often provides for more comprehensive evaluation. This chapter reviews a number of pain-related assessment measures, the constructs they purport to measure, and then briefly addresses issues related to normative data and symptom validity for each.

MMPI disability profile: the least known, most useful screen for psychopathology in chronic occupational spinal disorders

STUDY DESIGN

Prospective study on predicting psychopathology in chronic occupational spinal disorders (COSDs).

OBJECTIVE

To assess prevalence of specific profiles on the Minnesota Multiphasic Personality Inventory (MMPI) and their ability to predict psychopathology in a COSD cohort.

SUMMARY OF BACKGROUND DATA

In the relatively small number of COSDs that develop chronic pain and disability, the MMPI-2 has been an important part of the psychosocial assessment. Certain profiles have been thought to have a high prevalence in COSD. They have also been widely popularized as predicting certain treatment outcomes, and have often been used to "screen" surgical and rehabilitation candidates.

METHOD

Of an initial cohort of 1,489 consecutive COSD patients completing a valid prerehabilitation MMPI-2, 1,185 patients (79.6%) were classifiable into one of four MMPI profile groups. A new Disability Profile (DP) group was identified, which was the most common profile. Patients attended a 5- to 7-week interdisciplinary rehabilitation program. They completed a psychosocial assessment battery, and a Structured Clinical Interview for DSM IV diagnosis (SCID-I and II) was administered as the "gold standard" for defining psychopathology. One year postrehabilitation, a structured clinical interview assessed socioeconomic outcomes.

RESULTS

A previously unrecognized MMPI profile, now termed the DP, was found to have a prevalence of 53.2% of the whole group, and 66.9% of those with "classifiable" MMPI profiles in this large population of COSD patients. Only 6.9% of subjects had normal profiles (NP), while only 19.5% had profiles previously thought to occur commonly in this population. NP patients were twice as likely to retain work 1 year after treatment than the 3 abnormal MMPI groups combined. The DP group was 14 times more;1 likely to have an Axis I diagnosis (such as depression or anxiety) than the NP group, and was also almost 5 times more likely have an Axis II personality disorder diagnosis.

CONCLUSIONS

The prevalence of commonly cited MMPI profiles, often used for presurgical or chronic pain screening in this population, is relatively small. The prevalence of four or more elevations (DP), however, is large, representing two thirds of patients demonstrating any classifiable MMPI pattern. The DP group showed extremely high levels of associated psychopathology, which raises "red flags" to the surgeon likely to operate on such patients, or the interdisciplinary pain team. Screening COSD patients with the MMPI-2 may be effective in identifying psychopathology, but only if the physician is aware that the DP is commonplace and significant.

MMPI-2 subgroups in a sample of chronic pain patients

MMPI-2 scores of 307 female and 161 male chronic pain patients were analyzed by gender using a multivariate clustering method. Two subgroups were found for both sexes replicating previous results. The major subgroup corresponded to the classical "Conversion V" and the minor corresponded to the "Generally elevated" profile. The results also indicated a satisfactory internal consistency and a high discriminant validity of the Swedish version of the MMPI-2.

The Potential of Treatment Matching for Subgroups of Patients With Chronic Pain

A large and diverse number of treatments have been shown to be effective in reducing pain and other symptoms for a minority but statistically significant number of patients in different chronic pain syndromes. The means by which such different treatments achieve similar outcomes is not well understood. In this paper, the importance of considering patient heterogeneity for those who may be diagnosed with the same medical syndrome is discussed. The author suggests that the lack of satisfactory treatment outcomes for the treatments of chronic pain syndromes may be accounted for by the patient homogeneity myth--the assumption that all patients with the same medical diagnosis are similar on all important variables. The importance of subdividing (splitting) patients into meaningful groups is described. Studies presenting data on the identification of patient subgroups based on psychosocial and behavioral characteristics and the reliability and validity of this approach are presented. Some initial attempts to demonstrate the potential for matching treatments to patient subgroups are described.

A fourth empirically derived cluster of chronic pain patients based on the multidimensional pain inventory: evidence for repression within the dysfunctional group

The authors proposed that chronic pain patients with repressive defenses are not represented in current 3-cluster solutions of the Multidimensional Pain Inventory (MPI; R. D. Kerns, D. C. Turk, & T. E. Rudy, 1985) and that such a group can be distinguished by using a measure of defensiveness together with subscales of the MPI. They expected these patients to be described both by high defensiveness and by elevated pain and disability but minimal emotional distress. For 178 pain patients, hierarchical cluster analyses were performed on the MPI and Balanced Inventory of Desirable Responding (D. L. Paulhus, 1984). A 3-cluster solution replicated past findings in identifying dysfunctional, interpersonally distressed, and adaptive coper groups. A 4-cluster solution fit the data better, with a repressor group described by high pain, low activity and low distress emerging from the dysfunctional group. Profile analysis of validation measures showed that repressors scored comparably with dysfunctional patients on somatic symptoms of depression, pain severity, and perceived disability but significantly higher on these factors than the adaptive copers. Repressors scored comparably with adaptive copers on cognitive-affective symptoms of depression, anxiety, and anger but significantly lower on these variables than dysfunctional patients. Repressors also reported greater pain severity and perceived disability relative to their reports of negative affect, whereas dysfunctional and adaptive coper groups exhibited no such disparities. Without a measure of defensiveness, the MPI may misclassify a distinct group of patients as dysfunctional, but who readily endorse physical symptoms yet report low levels of emotional distress.

The Minnesota Multiphasic Personality Inventory and chronic pain: a conceptual analysis of a long-standing but complicated relationship

The Minnesota Multiphasic Personality Disorder (MMPI) and its successor, the MMPI-2, have a long-standing tradition in the assessment of patients with chronic pain. With the introduction of more narrowly defined and factor-analyzed pain inventories, however, the utility of the MMPI-2 for pain assessment has been brought into question. In this review, the relevant literature is carefully scrutinized from a conceptual and historical perspective. It is concluded that many of the (recent) criticisms are largely ungrounded. Rather than the test itself being at fault or of little utility in the field of pain assessment, it has simply been applied inappropriately (i.e., for determination of pain etiology or underlying personality structure "explaining" the chronic pain). In conclusion, it is suggested that the application of the MMPI-2 in the assessment of patients with chronic pain should correspond more closely to the original aims and psychometric properties of the tool--that is, for screening and the generation of hypotheses regarding comorbid psychopathology and personality features having the potential to complicate the treatment process. Guidelines for clinical interpretation of MMPI-2 profiles with regard to chronic pain are provided.

Cluster of MMPI personality profiles in chronic tension-type headache and predictable response to Fluoxetine

A study was made of the personality profiles of a sample of 51 patients with chronic tension-type headache (CTH) employing the Minnesota Multiphasic Personality Inventory (MMPI). Two clusters were obtained by applying multivariate classification techniques: cluster 1 (with elevations on scales Hy, D and Hs only) and cluster 2 (exhibiting elevations on most scales except Pd, Mf and Si). Fifty subjects without chronic pain or known psychiatric disorders, and extracted from the same social setting as the patients, completed the MMPI as a control group. Fluoxetine treatment was started in the CTH group, with follow-up over a 1-year period. Chi-squared analysis correlating the clusters obtained to different pain-related variables and epidemiological parameters revealed a significant association to sex only. There were no differences in therapeutic response between the two clusters. However, the patients belonging to the less perturbed cluster who exhibited profiles analogous to those of the control population showed significant improvement with respect to the global sample and their own cluster.

Validity of MMPI-2 profiles in chronic back pain patients: differences in path models of coping and somatization

OBJECTIVE

To show clinical utility and empirical validity of Minnesota Multiphasic Personality Inventory-2 (MMPI-2) chronic pain patient subgroups by identification of differential multivariate relationships across groups.

METHOD

This study used structural equation modeling to test cognitive coping strategies and somatization as mediator variables in path models with pain severity and depression used as exogenous (independent) variables and patient's activity level as the final endogenous (dependent) variable, across MMPI-2 profiles.

RESULTS

Hierarchical cluster analysis, performed on a sample of 569 chronic low back patients, resulted in four cluster profiles identifiable as those found in previous work with the MMPI-2 (within normal limits, V-type, neurotic triad, and depressed-pathological). Somatization mediated the relationship between depression and activity level for the neurotic triad group but not the other three groups. A positive linear relationship was found between somatization and depression for the within normal limits, neurotic triad, and depressed-pathological groups, whereas their linear association was negative for the V-type group. Cognitive coping strategies mediated the relationship between depression and activity level for the within normal limits group. In addition, cognitive coping was predictive of activity level for the within normal limits, V-type, and neurotic triad groups but not for the depressed-pathological group.

CONCLUSION

Consistent with previous cluster analytic studies, this study replicated four MMPI-2 cluster profile groups in chronic pain patients. These results have also shown that several multivariate relationships between variables are different across MMPI-2 groups, providing evidence for the validity for these MMPI-2 subgroups.

Effects of psychiatric consultation on medical consumption in medical outpatients with low back pain

A randomized, controlled, clinical trial (N = 104) was conducted to test the hypothesis that a protocol of collaboration and communication between neurologist and general practitioner, sustained with psychiatric consultation, would reduce medical consumption (especially of diagnostic procedures and medication) in medical outpatients with low back pain. The intervention was designed at the health care provider level; the psychiatrist did not see or examine the patient. A reduction in medical consumption could not be demonstrated. As there was a great variation in adherence to the protocol in the experimental group, this might explain the lack of hypothesized effects. Full implementation of the protocol seemed to imply a lower number of major surgery operations, but this effect disappeared after excluding cases with a diagnosed hernia. The possibilities for consumption reduction for three post hoc-defined patient categories (cases with hernia and chronic and nonchronic cases) and the consequences for patient selection and treatment intensity are discussed. It is concluded that the target groups should be more narrowly defined, and that a more intensive intervention might prove to be more effective.

Typological description of the chronic low back-pain syndrome using the Million Behavioral Health Inventory

The Million Behavioral Health Inventory was administered to 67 individuals with chronic low back pain who participated in a rehabilitation program on pain and stress. Base-rate scores from the 20 scales, submitted to cluster analysis using Ward's minimum variance method, gave four subtypes. Discriminant analysis accurately classified 98.51% of the sample. The discriminant functions were interpreted and labeled as I: somatic distress and II: inhibition-dependency. The groups were then examined for differences and similarities with respect to those dimensions, and the groups were labeled I-introversive, II-denial/minimizers, III-conformers, and IV-severe psychophysiological reaction. The types were next compared across medical, vocational, and demographic data. With the exception of sex of subject, no other data were significant by type. The results were discussed, and recommendations for research were made.

Relationship between the Millon Behavioral Health Inventory and the minnesota multiphasic personality inventory (MMPI) in low-back pain patients

The Millon Behavioral Health Inventory (MBHI) is being used with increasing frequency for the assessment of chronic pain, although there is a relative lack of evidence as to its utility, and prior studies have not examined low back pain. This investigation compared the MBHI to the MMPI in a sample of low-back pain patients and analyzed subgroups of pain patients based upon their MBHI responses. Subjects were 60 patients who had been admitted to outpatient multidisciplinary pain clinics of two Chicago-area hospitals. Patients completed both the MMPI and the MBHI and provided demographic information. Results of correlational analyses indicated strong relationships between the MBHI psychogenic attitude, psychosomatic correlate, and prognostic index scales and the validity scales of the MMPI. The MBHI Pain Treatment Responsivity scale (PP) correlated with 16 of the other 19 MBHI scales. PP did not demonstrate specificity with low back pain patients. The results of both the scale comparisons and the exploratory two-group cluster subgroup analysis support the notion that responses to the MBHI are largely affected by the respondent's tendency to deny psychopathology or to admit emotional distress.

Subtypes of psychosocial adjustment to breast cancer

By means of cluster analytic techniques, four subtypes of psychosocial adjustment were identified in a sample of 122 breast cancer patients who completed the Psychosocial Adjustment to Illness Scale. Internal consistency and internal validity of the derived typology were suggested by the finding that two different hierarchical agglomerative clustering methods (average linkage between groups, Ward's) produced similar solutions. Three of the derived subtypes reported normal affect levels but different patterns of relative strengths and dysfunctions, while the fourth subtype appeared to be highly distressed and globally maladjusted. External validation was demonstrated by differentiating the subtypes on variables of negative affect, avoidance coping, and fighting spirit. The clinical and heuristic implications of these findings are discussed. The findings highlight the need for comprehensive assessment of psychosocial functioning of cancer patients. They demonstrate that even non-emotionally distressed patients can have very different profiles of adjustment and may benefit from correspondingly individually tailored psychosocial interventions.

Minnesota Multiphasic Personality Inventory (MMPI) cluster groups among chronically ill patients: relationship to illness adjustment and treatment outcome

Cluster analysis of the MMPI has been utilized widely in the chronic low back pain literature to try to identify reliable patient subtypes predictive of treatment outcome. We extended this methodology to patients with heterogeneous chronic medical conditions by replicating prototypic MMPI cluster group profiles and by relating cluster groups to clinical baseline and outcome data. Subjects were two independent samples (n = 254 and n = 263) of chronically ill patients admitted to an inpatient medicine/psychiatry unit. Using a four-cluster solution, similar cluster profile groups were replicated in both samples. Consistent differences emerged between cluster groups on functional impairment, psychiatric diagnoses, depression, and psychosomatic symptoms. Cluster group membership also predicted changes in functional impairment and depression six months after treatment. Results are discussed in terms of similarities between chronic low back pain and chronic illness and tailoring treatment to different patient types.

An empirical evaluation of multidimensional clinical outcome in chronic low back pain patients

Individuals with persisting pain often present a constellation of symptoms that includes pain, health-related impairment and dysphoric mood. It is now widely accepted that comprehensive assessment must address each of these dimensions. Despite recognition of the value of multidimensional assessment, no empirical efforts have validated the construct of a multidimensional clinical outcome presentation based on the dimensions of pain, impairment and dysphoric mood. We employed cluster analytic procedures on standard measures of pain, impairment and depression in chronic low back pain (CLBP) patients (n = 96) attending a general orthopedic clinic in order to empirically characterize multidimensional clinical outcomes. Results indicated that 3 groups could be identified reliably: (1) 'Chronic Pain Syndrome' (n = 25; high levels of pain, impairment and depression), (2) 'Positive Adaptation to Pain' (n = 24; high levels of pain with low levels of impairment and depression) and (3) 'Good Pain Control' (n = 47; low levels of pain, impairment and depression). The reliability of this cluster solution was supported by several tests of internal consistency. Discriminability of the clusters was examined across both the outcome measures themselves and several additional independent variables. The cluster solution was then cross-validated in an independent sample of pain clinic CLBP patients (n = 180) to test its generalizability. Finally the stability of the cluster dimensions over time was tested by re-assessing 36 CLBP patients 6 months after they initially were characterized into 1 of the 3 outcome groups on the same measures. MANOVA results indicated that the outcome groups were differentiated statistically across assessments. The multiple outcome measures did not change significantly across time, nor did the outcome groups change differentially across time on these measures. We conclude that the outcome dimensions of pain, impairment and depression are relatively stable phenomena that differentially describe CLBP patients.

Psychological factors in the long-term prognosis of chronic low back pain patients

This study appraised the significance of psychological factors in the long-term prognosis of patients with chronic low back pain (LBP). The MMPI ratings of 80 long-term sick-listed LBP patients were set in relation to their disability pension status 6 to 12 years later. The number of elevated scales, in combination with the level of certain scales (HS and HY), proved to be a better predictor than profile patterns, advocated in some studies. Moreover, in a review of prognostic studies, the HS and HY scales appeared most frequently as significant predictors. That result also was confirmed in this study of long-term prediction of overall functional level. The results are discussed in relation to the concepts of pain-fear and sick role.

Patterns of normal personality structure among chronic pain patients

Fifty-nine chronic pain patients satisfying one of four previously identified pain group classifications were evaluated using the NEO Personality Inventory (NEO-PI), a standardized measure of normal adult personality structure. Minnesota Multiphasic Personality Inventory (MMPI) pain subgroups differed with respect to level of NEO-PI Neuroticism. In particular, emotionally overwhelmed pain patients as defined by multiple MMPI scale elevations had higher NEO-PI Neuroticism scores. Post hoc analyses revealed higher levels of depression, anxiety, vulnerability, and hostility in emotionally overwhelmed subjects. None of the remaining groups differed from each other on NEO-PI Neuroticism. Additionally, none of the other NEO-PI domains discriminated pain subgroups. NEO-PI profiles for pain patients (except for Neuroticism in emotionally overwhelmed patients) yielded t scores in the average range, suggesting that chronic pain patients present with a relatively normal underlying personality structure.

Health locus of control, gender differences and adjustment to persistent pain

Locus of control (LOC) beliefs, long thought important in adjustment to persistent pain, were studied among 160 subjects (67 males and 93 females) referred to a comprehensive pain rehabilitation program. The subscale structure of the Multidimensional Health Locus of Control (MHLC) was factorially replicated in our sample. Three unique MHLC profile clusters were identified for both males and females. Among men, cluster assignment was related to age only. The younger male patients reported a stronger internal attributional style. Older male patients relied more heavily on both chance and powerful other factors. Among women, cluster assignment was related to the use of coping strategies. For example, patients with high internal scores only, reflecting a strong internal orientation towards self-management of health care needs, were more likely to utilize Information-Seeking, Self-Blame, and Threat Minimization coping strategies than patients with high scores on both the Internal and Powerful Other factors. It appears that the presence of both Internal and Powerful Other health attributional styles is associated with less frequent use of cognitive self-management techniques. In understanding the LOC scores it is important to rely on pattern analysis of scores. Implications for clinical treatment are discussed.

Behavioral assessment of low back pain: identification of pain behavior subgroups

A hierarchical cluster analysis procedure was used to identify homogeneous subgroups of low back pain patients who show similar pain behavior patterns during a videotaped behavior sample. Subjects, 106 chronic low back pain patients, were divided into 2 samples. The cluster analysis procedure identified 4 very similar subgroups in sample 1 and sample 2. The first subgroup showed a low or moderate level of guarding and a low level of other pain behaviors. The second subgroup showed a high level of guarding and a moderate level of rubbing of the painful area. The third subgroup showed high levels of guarding and moderate levels of bracing and rubbing. The fourth subgroup exhibited a very high level of rubbing and moderate levels of bracing. The subgroups identified may require somewhat different approaches to pain assessment and treatment. The results of this study suggest that the variability in pain behavior evident in low back pain populations may be due to the fact that within these populations there are homogeneous subgroups of patients who differ in the motor pain behaviors they display.

A comparison of clinically and empirically derived MMPI groupings in low back pain patients

Two methods of grouping MMPI scales were employed with 271 patients who were attending a Back Clinic at a university hospital. One method involved clinically derived decision rules, while the other utilized a multivariate clustering procedure. Five subgroups of patients were identified with each method. Both procedures showed subgroup differences in terms of pretreatment pain intensity, activity limitations, McGill affective scores, and physician ratings of degree of functional impairment; the elevated subgroups showed more dysfunction. Follow-up comparisons revealed that, regardless of the classification procedure used, the elevated subgroups showed poorer response to treatment. Only the clinically derived groupings had significant differences in chronicity of pain and total scores on the McGill Pain Questionnaire. On the other hand, only the statistically based system showed significant differences on age and marital status. Furthermore, only the statistical system identified subgroups of male patients that predicted followup pain intensity decreases in a manner that indicated clinical usefulness.

Sociodemographic and clinical correlates of P-A-I-N

Rules to clarify MMPI profiles into four types (P-A-I-N) observed commonly in chronic pain patient populations were tested to determine if resulting average profiles matched computer-algorithm derived prototypes. Close matches were obtained, suggesting that the rules could be used in the ordinary pain clinic setting to look for type x treatment interactions. Seven clinical variables differentiated the types. Type P and N were most different from one another, and Types A and I represented poles on an optimism-pessimism dimension.

The detection of psychiatric illness and psychological handicaps in a British pain clinic population

Ninety-seven successive patients attending the Newcastle Pain Relief Clinic completed a battery of psychiatric, psychological and pain questionnaires, and an extensive personal information form. All patients were seen by a physician who evaluated the extent of the pain arising from physical, psychiatric and psychological causes, and by a psychiatrist, who administered a structured interview schedule. Thirty-two percent of the patients had sufficient symptoms to be classified as psychiatric cases on the Present State Examination (PSE), a further 22% had minor neurotic symptoms and features of illness behaviour, 35% were categorized as organic, and 11% were unclassified. The Leeds General Depression Scale for Depression and Anxiety and the Beck Depression Inventory were the most effective of the psychiatric questionnaires used in separating the psychiatric patients from the remainder, and can be recommended as screening instruments for psychiatric illness in this population. Factors associated with a psychiatric diagnosis included female sex, larger number of present medications, greater reduction in activities compared to the period before the pain developed and increasing subjective pain from the onset of this.

Empirically derived Symptom Checklist 90 subgroups of chronic pain patients: a cluster analysis

Four hundred fifty-three chronic pain patients completed a Symptom Checklist 90 (SCL-90) and a comprehensive pain evaluation questionnaire. All patients were evaluated by a physician and rated on degree of pain pathology and pain behavior. The SCL-90 data were analyzed using two clustering procedures and replicated over two similar samples. Three distinct profiles emerged and represented high, medium, and low scores on the SCL-90. No differences were found between subgroups on demographic characteristics, compensation status, pain duration, or pain ratings. Patients in the high-profile subgroup showed the most emotional distress, reported that their pain interfered the most with all activities, and were most often judged to have high pain behavior. Little evidence was found to support a "denial" profile, as previously predicted. Further support was found for using the SCL-90 in assessing chronic pain patients.

Pain-patient Minnesota Multiphasic Personality Inventory (MMPI) subgroups: evaluation of long-term treatment outcome

Six hundred thirty-five chronic pain patients completed the Minnesota Multiphasic Personality Inventory (MMPI) prior to participation in a multidisciplinary inpatient pain treatment program. Three male and four female MMPI subgroups were identified by means of cluster analyses for each of two samples. Pretreatment and long-term follow-up differences were then examined among the MMPI subgroups. Results indicated that the subgroups identified in the present study closely resembled each other and those previously reported in the literature. However, at long-term follow-up only a few outcome differences were identified among male subgroups, while no differences were found among female subgroups. Possible explanations for no differential treatment outcomes among the MMPI subgroups are discussed.

MMPI profiles of chronic pain patients: some methodological considerations that concern clusters and descriptors

Significant methodological problems in the identification of MMPI subtypes in patients with chronic pain are addressed, and some illustrative data are offered. Particular attention is given to the selection of MMPI predictor data, determination of membership in a subtype, and the selection of nontest descriptors. Some recommendations for improving the predictive validity of statements about the personality and behavioral characteristics of patients with chronic pain from the MMPI are made.

P-A-I-N: a four-cluster MMPI typology for chronic pain

A 4-cluster empirically derived MMPI typology for chronic pain sufferers has been demonstrated by combining the results of 10 investigative teams. These MMPI 'types' have been labeled P-A-I-N and appear to have important clinical and demographic correlates. Type P is the most 'psychopathological' looking as nearly all scales are usually elevated. Type P patients are extreme in their claims of physical illness, psychological distress and social maladaptation. Demographic correlates include poor education, high rates of unemployment, and limited household income. Type A is defined by a 'conversion V' on the 'neurotic' triad scales. It has no unique correlates. Type I has elevations on all of the neurotic triad scales and on no others. Type I patients appear to be the most physically infirm with multiple surgeries and hospitalizations. They may not improve physical status with treatment, but appear to benefit psychologically. Type N profiles are 'normal' in that no scale, except perhaps scale K, is often elevated. Type N patients are moderate in their claims of ill health, often are better educated and employed, and appear to respond well to treatment. Classification rules have been proposed to allow patient-typing without a computer. Use of these rules should allow programmatic research into treatment/type interactions even in the ordinary clinical setting. The typology appears well enough established to allow for prospective studies to test theoretical hypotheses drawn from the literature base.

The derivation and validation of six Multidimensional Health Locus of Control Scale clusters

The purpose of this research was twofold. First, to determine if eight Multidimensional Health Locus of Control types hypothesized by Wallston and Wallston (1982) existed. Second, to assess the reliability, validity, and clinical utility of the control types. Two investigations were conducted based on research procedures designed for the discovery of clusters. The first study involved the derivation and replication of MHLC clusters with a sample of 400 healthy undergraduate men and women. The results of Study 1 suggested the existence of six MHLC clusters: pure internal; double external; pure chance; yea sayer; nay sayer, and believer in control. The sample for Study 2 consisted of 90 female undergraduate and graduate students. The six-cluster solution was replicated in this second cross-validation study. Furthermore, construct validity of the clusters was established through computer simulation. Finally, relationships between clusters and the Krantz Health Opinion Survey suggested the clusters have a theoretical as well as a empirical foundation. The implications of these findings for clinical practice, for the development of the health locus of control construct, and for future research on the MHLC Scale are discussed.

Cluster analysis of Minnesota Multiphasic Personality Inventory (MMPI) profiles in a chronic headache population

Minnesota Multiphasic Personality Inventory (MMPI) profiles from 123 chronic headache sufferers (48 migraine, 47 muscle contraction, 28 mixed) were cluster analyzed in order to replicate results from other pain populations and to examine differences among the clusters along headache parameters and demographic variables. Four homogeneous subgroups were obtained and a multiple discriminant analysis was successful in correctly classifying 97.56% of the headache subjects. Obtained clusters closely resembled those evidenced in previous pain research. Results indicated that diagnostic category was unrelated to cluster membership; however, frequency of headache occurrence was a differentiating factor. Implications for headache treatment and future research are discussed.

The MMPI and psychological factors in chronic low back pain: a review

The Minnesota Multiphasic Personality Inventory (MMPI) is widely used in the psychological assessment of patients with chronic low back pain (LBP). Patients' profiles have been used in a number of ways: in attempts to discriminate between cases; as predictors of both medical treatment and pain management program outcomes; and in attempts to assess degree of disability. Studies reviewed here indicate that the concept of psychological etiology of chronic LBP, despite widespread use, has failed to differentiate patients and to reliably predict response to specific treatment. A promising alternative approach has emerged in recent years: profile distinctions between different types of psychological response to chronic LBP. These subgroups are associated with different pain-related behaviors and may show differential response to various treatments, although further work remains to be done to specify the relationships more precisely. Methodological difficulties that continue to appear in the literature are addressed and recommendations for further developments in the use of the MMPI with this patient population are made.

Cognitive and somatic item response pattern of pain patients, psychiatric patients, and hospital employees

Standard psychological tests generally provide a single global score that reflects multidimensional constructs, such as depression and anxiety. This single score, however, integrates a range of item contents, including cognitive/affective, somatic, and behavioral characteristics of these multidimensional constructs. The present study was designed to compare the pattern of item endorsement among chronic pain patients (N = 50), psychiatric inpatients (N = 50), and hospital employees (N = 50) on the SCL-90-R (Derogatis, Rickels, & Rock, 1976). Pain patients reported the highest SCL-90 scale level of Somatization, while the psychiatric inpatients reported the highest level of Anxiety and Depression. Additionally, the within-scale pattern of item responses on the Anxiety and Depression scales differed among groups. Although psychiatric inpatients endorsed equivalent levels of somatic and cognitive items, the pain patients' reports of psychological distress were limited primarily to somatic signs of anxiety and depression. Thus, the interpretation of pain patients' psychological profiles and subsequent treatment recommendations may be inappropriate if based on normative data obtained from psychiatric and/or normal populations.

A multimethod evaluation of the MMPI in a chronic pain patient sample

Previous reports on the use of the MMPI with chronic pain patients have produced a variety of results. No single configural feature or scale identifies the chronic pain patient, regardless of the origin or verifiability of the etiology of the pain. Elevations on the neurotic triad occur frequently, but do not have the specificity of more recently reported chronic-pain subtypes on the MMPI. The present nonexperimental study (N = 72) provides an example of a multi-method analysis of a carefully selected sample of chronic pain patients without physical findings. Blind clinical analysis, simple two-point code aggregation, and multivariate profile methods were used and produced similar subtypes of the sample and understanding of the data. The obtained sample subtypes were similar to those found in other studies and were hypothesized to be related to each other along an underlying continuum of what might be depression. Further research is needed to facilitate understanding of the causation of chronic pain of obscure origin.

Empirically derived pain-patient MMPI subgroups: prediction of treatment outcome

Fifty-seven male chronic pain patients admitted to an inpatient multimodal pain treatment program at a Midwestern Veterans Administration hospital completed the MMPI, Profile of Mood States (POMS), Tennessee Self-Concept Scale (TSCS), Rathus Assertiveness Schedule (RAS), activity diaries, and an extensive pain questionnaire. All patients were assessed both before and after treatment, and most also were assessed 2-5 months prior to treatment. No significant changes occurred during the baseline period, but significant improvements were evident at posttreatment on most variables: MMPI, POMS, TSCS, RAS, pain severity, sexual functioning, and activity diaries. MMPI subgroup membership, based on a hierarchical cluster analysis in a larger sample, was not predictive of differential treatment outcome. Possible reasons for comparable treatment gains among these subgroups, which previously have been shown to differ on many psychological and behavioral factors, are discussed.

Chronic pain: an investigation of assessment instruments

The present study investigated professional perceptions of utility and the actual predictive validity of the MMPI and the Millon Behavioral Health Inventory (MBHI) in a sample of chronic pain patients. Overall results indicated statistical preference for the MBHI on nine of the ten dimensions of clinical utility. In addition, patient prognosis was rated as more favorable based upon the MBHI report. The differences noted between raters of different professional capacities were hypothesized to reflect differences in familiarity with assessing pain patients from a psychological perspective. However, over the course of a one year follow-up, neither instrument significantly predicted such behaviors as hospital usage, medication usage or general status. Rather, the psychologists who wrote the original consultations (which were not rated for utility) were more accurate than either the MMPI or the MBHI in predicting these health related behaviors.

A replicated cluster solution in a heterogeneous psychiatric population

Two cluster analytic studies were conducted to explore groupings of patients in a heterogeneous psychiatric population. Study 1 (N = 170) provided support for the derivation by a cluster analytic method of five meaningful clinical groups of patients. Based on an extensively modified inventory and a different population of psychiatric patients, results from Study 2 (N = 224) provided further indication of the stability of the clusters. It was concluded that cluster analysis may provide a viable method as a classification system for a heterogeneous population of psychiatric inpatients.